Menu Sign In Contact FAQ
Banner
Welcome to our forums

Legacy autopilot questions

achimha wrote:

When you order an S-TEC autopilot, you specify the airframe. The factory then adjusts the pots based on data they have.

And you are not allowed to sell it on, even to the owner of exactly the same airframe. Vultures.

EGKB Biggin Hill

which describes how to adjust the parameters, but not really what they do.

Hmmm… interesting reading! I will have another go:

Pitch Trim – sets the pitch sensor null (for level flight i.e. VS=0) BUT at what speed? The pitch varies with speed… most planes cruise about 2.5 deg UP but it varies
Roll Trim – sets the roll angle sensor null (wings level) – this will not be speed dependent
Roll Gradient – sounds like the limit for roll angle; they are using 22 deg which is about Rate 1 for a reasonably fast piston plane
Roll Gain – seems to be a cruise speed version of Roll FB
Pitch Gain – as above – the prime case of porpoising is when this is too high
Roll FB – seems to set the derivative term for the loop gain for roll
Pitch FB – as above, for pitch
Intercept – sounds like the target for the rate at which the VOR radial is changing (set to achieve 45 degrees)
G/A Pitch up – self explanatory, sets the angle in the G/A mode
Pitch Comm Max – appears to set the maximum pitch the AP will command under any circumstances

I am impressed that they use high quality sealed multiturn pots. These might actually last a good few decades.

This bit

is not trivial since these are roughly 20 turn pots and you can’t see what they are set to, plus they have no “feel-able” end stops (there is a clutch which slips at the stop) so the only way to establish position is to note how many turns you have done (and which way) and then you can go back to it.

Whoever designed this must have enjoyed it. This is from Apollo era; no wonder America got to the moon, and then established a world monopoly in so many fields.

And you are not allowed to sell it on, even to the owner of exactly the same airframe. Vultures.

Like many aviation practices (e.g. STCs limited to dealers) I am sure this is illegal in Europe, but being below the radar it will just carry on. However, I have just taken delivery of the full TKS system and the STC for that is airframe and named installer specific. CAV assure me they are happy to change the installer name if required, but if they didn’t then you have scrap.

Administrator
Shoreham EGKA, United Kingdom

That is all very helpful, thank you!

Peter wrote:

Roll FB – seems to set the derivative term for the loop gain for roll

What does that mean in practice?

EGKB Biggin Hill

It is all to do with achieving relative stability.

If you google on e.g. control system theory (a vast subject, developed mathematically about 100 years ago onwards) you find the basic principles. If you want to control something, say the speed of a car by varying engine power, you have a range of behaviour which starts at very sluggish, and ends with extreme overshoots/oscillation. In between the two you get, at the simplest “classical” case, what is called critical damping where you get just a small overshoot. In autopilots they will over-damp slightly because you don’t want any overshoot so you start to level off before the target.

This challenge is unavoidable; there is no way to “just control something” if there is feedback involved.

Nowadays most control systems are in software so they add various tweaks beyond the classical PID stuff. These are especially desirable where there are time lags (e.g. propagation time through the insulation in an oven; this plays havoc with control). Also for many years one has had self-learning (“self tuning”) algorithms. But yours is analogue and implementing clever features requires more and more components and in 1970 or so it really would have been a large number of components.

Does yours have an IAS input?

Administrator
Shoreham EGKA, United Kingdom

Peter wrote:

Does yours have an IAS input?

No. It is very basic. The system was originally supplied with an AI and HSI, but I have substituted an Aspen. It works well, but the dynamics are very different, hence the need to tune, even though it has just come back from Autopilots Central all correctly set up per the manual. The intercept is especially rubbish, so I have to steer it onto the ILS and then press APPR, but even then it is pretty much pants, until I wind that intercept angle pretty much to minimum.

EGKB Biggin Hill

OK; the Aspen is very likely to have a gain adjustment on its FCS (autopilot) output; probably more than one. For example the millivolts/degree for the heading bug being off “straight ahead” should be adjustable. On my Sandel SN3500 all this can be tweaked. Whether these adjustments are accessible I have no idea, but the EFD1000 IM should have the info.

Sounds like the Aspen is outputting way too much, resulting in way too high a loop gain.

Administrator
Shoreham EGKA, United Kingdom

I believe that there is a digital to analogue converter between the two.

EGKB Biggin Hill

Unfortunately, there is not an approved link from the Aspen AI to the autopilot, so I have had to keep the legacy AI, but it’s no bad thing to have a suction AI as well as the electric one. More unfortunately, neither AI gets FD input…I don’t really understand why not.

So the information from the Aspen is only from HDG bug, GPSS and HSI OBS displacement. My experience is that the intercept angle has to be reduced to an absolute minimum to get it to fly a coupled approach properly (otherwise it diverges) but I have been informed in PM by an avionics shop that the DAC can be tuned by replacing resistors to ensure that the voltage received by the a/p is correct, without trimming the a/p pot. I am not sure why this is advantageous to just tuning it at the pot.

EGKB Biggin Hill

There are still quite a few of us with original Bendix autopilots (sometimes rebranded as Altimatic, Cessna or whatever) and it has been increasingly difficult to get them serviced.

Everyone recommends Autopilot Central, but for the last four to six weeks they have just stopped answering emails. I even wonder if they have gone bust.

But today I had to be in Straubing anyway and asked them to look at my Altimatic V.

There is a guy here called Martin who has specialised in nothing else for over 20 years, but is still relatively young (compared to the walking dead who are rest of the legacy autopilot business )

He got in the aircraft, held onto the yoke and said “there’s oil in the roll servo clutch, we need to up the voltage to the servo and I think the connector is corroded” (all this from cockpit). He pulled the servo and clutch and fixed them in his workshop and resoldered a plug. He then jumped back in, fiddled for about 60 seconds and said “the wrong transformer has been fitted between the ACU2 and the autopilot, never mind, we’ll just tune it out.”

He then went tweak, tweak, tweak, tweak on the pots (he looked more like an organist than an engineer) and said “that’ll be perfect now, but would you like to fly to be sure?” So we did, and it was!

It was like working with an artist. (And a really nice bloke as well.). All for less than €500.

They have complete repair facilities for all parts of the system (I am going to send our pitch trim gyro there instead of Autopilots Central, since they have gone completely silent on me.)

So, seriously, anyone with a legacy autopilot problem, take a holiday in Bavaria. You won’t regret it!

EGKB Biggin Hill

Nice one Timothy! Glad they found out what it was, so no magic but simply a guy who knows what he is doing.

I remember a similar case with someone who had a Brittain AP installed and got an Aspen and could not get it to work with that one. All the knowalls in the fora went “you can’t connect an Aspen to a Brittain AP” but then he was directed towards the manufacturer’s website who sorted him out within a few days. It was not the Aspen but the Brittain AP had been previously misadjusted to compensate for a DG whose output was substandard. So from the Aspen converter he got the correct output which then proved way too high and sent the plane dancing around the heading…

LSZH(work) LSZF (GA base), Switzerland
Sign in to add your message

Back to Top